Since the databases that banks use for analysis of cardholders’ repayment behaviours are usually
large and complicated, and the extant classification techniques hardly offer 100% correct
classification accuracy so as to possibly incur a considerable loss associated with type II errors, the
prediction of cardholders’ future payment behaviours has been still referred to as a difficult task in the
credit industry.
This paper proposes a two-stage cardholder behavioural scoring model, with merits of artificial
neural networks (ANNSs) and data envelopment analysis (DEA), which not only enables banks to
verify the ANNSs predicted results of each cardholder’s future repayment behaviour as well as to
identify creditworthy cardholders who is profitable with low risks, but also provides guidelines to
improve contributions of each inefficient cardholder for card issuer profitability.
關聯:
International Journal of Advancements in Computing Technology 3(2)
